Biologic and prognostic significance of the presence of more than two mu heavy-chain genes in childhood acute lymphoblastic leukemia of B precursor cell origin

We examined the arrangement of the mu heavy-chain immunoglobulin (Ig) genes in the leukemic blast cell DNA of 93 children with acute lymphoblastic leukemia (ALL). All cases met morphologic and cytochemical criteria for ALL, lacked detectable T cell surface antigens, and expressed HLA-DR (Ia) antigens. Eighty-three of the 93 patients (89%) were positive for the common acute lymphoblastic leukemia antigen (CALLA), and 20 of 91 (22%) tested had detectable cytoplasmic immunoglobulin. As expected, the heavy-chain lg gene was rearranged in all cases, and the pattern of rearrangements was variable; 23 had one allele rearranged and one in the germ line configuration; 15 had one rearranged and one deleted; and 37 had two rearranged. Unexpectedly, in 18 patients the presence of more than two mu gene-hybridizing bands was detected. Combinations of enzymes and heavy-chain gene probes were used to confirm that the extra bands were not the result of underdigestion of the DNA or DNA restriction site polymorphism. In eight of the 18 patients, we identified an extra chromosome 14 as a possible cause of the extra bands' hybridizing to the mu heavy-chain constant-region probe. In the remaining ten patients, the presence of three or four bands hybridizing with the mu probe suggests the presence of two populations of leukemic cells that may have arisen either by separate leukemic transformation events or by clonal evolution of one clone into two related lines. Although preliminary (2-year follow-up), our data suggest that childhood ALL of B lineage with more than two mu heavy-chain genes, but without extra copies of chromosome 14, may be more resistant to therapy.     

Multiple DNA fragment polymorphisms associated with immunoglobulin mu chain switch-like regions in man.

DNA probes containing the switch region (S) associated with the human immunoglobulin heavy chain mu gene were used to investigate polymorphisms in the germ-line human DNA. Six polymorphisms, detected by a single restriction enzyme (Sst I) are described. Linkage studies in 29 families show that five of the six polymorphisms, although relatively unassociated in random individuals, segregate in complete linkage one to the other and to Gm allotypes (markers on the heavy chain of IgG), while the sixth segregates independently. Altogether, when one considers the DNA markers at the five closely linked loci and the IgG1 and IgG3 heavy chain allotypes, 33 different haplotypes have been described; of these, 28 are detected by the DNA polymorphism alone. Study of 158-187 random haplotypes showed strong linkage disequilibrium only between one DNA polymorphism (Sst A) and Gm. Of the polymorphic Sst I loci, one, Sst E [associated with 2.2- to 2.7-kilobase (kb) fragments], is included in the mu chain S region (S mu); another, Sst A (6.8-7.4 kb), must be very close to the gamma 1-gamma 3 chain gene cluster. Based on studies of an IgE human myeloma, a third polymorphism, Sst C (4.8-5.5 kb), should map 3' of the active epsilon chain gene. An Sst I restriction enzyme map of phage clones carrying the two alpha chain genes indicates that Sst A and Sst C loci probably overlap with the alpha 1 and alpha 2 S regions, respectively. Both deletion/duplications and point mutations were detected.     

Structural differences in a single gene encoding the V kappa Ser group of light chains explain the existence of two mouse light-chain genetic markers.

Two phenotypic markers of mouse immunoglobulin kappa light chains, the IB-peptide marker and the Ef1a isoelectric focusing marker, are expressed by the C58/J, AKR/J, RF/J, and PL/J strains (called expressor strains) but not by BALB/c and most inbred strains. Expression is linked to the kappa light-chain locus and the Lyt-2/Lyt-3 genes on chromosome 6. Light chains bearing these markers belong to a group of variable region kappa chain (V kappa) regions called V kappa Ser, which has a serine amino terminus and a framework 1 region not observed to date among BALB/c light chains. Southern hybridization of genomic DNA with a V kappa Ser-specific cDNA probe has demonstrated a single strongly hybridizing DNA fragment in all strains of mice tested. Characteristic restriction enzyme polymorphisms define the V kappa Ser alleles of expressor (Igk-VSera) and nonexpressor (Igk-VSerb) strains. In the present study, the unrearranged V kappa Ser gene and its flanking regions from an expressor (C.C58) and nonexpressor (BALB/c) strain have been cloned and their nucleotide sequences determined. The C.C58 V kappa Ser gene isolated (the Igk-VSera allele) was shown to code for the two phenotypic markers described. While the nucleotide sequence of the BALB/c coding region (the Igk-VSerb allele) shows 97% identity with the C.C58 gene, single nucleotide substitutions lead to structural changes in the encoded protein which render it IB-negative and Ef1a-negative. These differences alone can explain the failure of strains containing the BALB/c allele to express these kappa-chain phenotypic markers. Also, the BALB/c gene contains a single substitution in a conserved octamer sequence approximately equal to 100 nucleotides upstream of the coding region, which could affect its expression. Finally, the C.C58 allele contains a BAM5/R repetitive DNA element approximately equal to 1200 nucleotides upstream of the coding regions that is not present in BALB/c. This element gives rise to the EcoRI and BamHI restriction enzyme polymorphisms, which distinguish the Igk-VSera and Igk-VSerb alleles.     

P450XXI (steroid 21-hydroxylase) gene deletions are not found in family studies of congenital adrenal hyperplasia.

Congenital adrenal hyperplasia (CAH) is a common genetic disorder due to defective 21-hydroxylation of steroid hormones. The human P450XXIA2 gene encodes cytochrome P450c21 [steroid 21-monooxygenase (steroid 21-hydroxylase), EC 1.14.99.10], which mediates 21-hydroxylation. The P450XXIA2 gene may be distinguished from the duplicated P450XXIA1 pseudogene by cleavage with the restriction endonuclease Taq I, with the XXIA2 gene characterized by a 3.7-kilobase (kb) fragment and the XXIA1 pseudogene characterized by a 3.2-kb fragment. Restriction endonuclease mapping by several laboratories has suggested that deletion of the P450XXIA2 gene occurs in about 25% of patients with CAH, as their genomic DNA lacks detectable 3.7-kb Taq I fragments. We have cloned human P450c21 cDNA and used it to study genomic DNA prepared from 51 persons in 10 families, each of which includes 2 or more persons with CAH. After Taq I digestion, apparent deletions are seen in 7 of the 20 alleles of the probands; using EcoRI, apparent deletions are seen in 9 of the 20 alleles. However, the apparently deleted alleles seen with Taq I do not coincide with those seen with EcoRI. Furthermore, studies with Bgl II, EcoRI, Kpn I, and Xba I yield normal patterns with at least two enzymes in all cases. Since all probands yielded normal patterns with at least two of the five enzymes used, we conclude that the P450XXIA2 gene "deletions" widely reported in CAH patients probably represent gene conversions, unequal crossovers, or polymorphisms rather than simple gene deletions.     

Characterization of frequent deletions causing steroid 21-hydroxylase deficiency.

Steroid 21-hydroxylase deficiency is caused by mutations in the CYP21B gene. This gene and a highly homologous pseudogene, CYP21A, alternate with the C4A and C4B genes encoding the fourth component of complement. Classical deficiency alleles are frequently caused by deletions of CYP21B or by gene conversions that transfer deleterious mutations from the CYP21A pseudogene to CYP21B. Gene conversions involving restriction enzyme sites that distinguish CYP21A [e.g., 3.2-kilobase (kb) Taq I fragment] and CYP21B (3.7-kb Taq I fragment) might be confused with actual deletions of CYP21B. To determine the incidence of this type of gene conversion, 15 chromosomes (in 13 families) with absent 3.7-kb Taq I fragments were studied. When hybridized with a 21-hydroxylase probe, all of these chromosomes were associated with absent 2.9-kb Kpn I fragments, 14 of 15 were associated with absent 2.4-kb Bgl II/EcoRI fragments, and 13 of 15 were associated with absent 10-kb Bgl II/EcoRI and 12-kb EcoRI fragments. Thirteen of 15 chromsomes had absent 6.0- or 5.4-kb Taq I fragments when hybridized with a C4 probe. Thus, 2 of 15 chromosomes do not carry deletions and may represent gene conversions; 13 of 15 chromosomes studied have a deletion of approximately equal to 30 kb, leaving behind the C4A gene and a single CYP21A-like gene. Hybridization with specific oligonucleotide probes showed that in all 13 cases this remaining CYP21 gene carried an 8-base-pair deletion, typical of CYP21A, that prevents synthesis of a functional protein. Thus, gene conversions are rarely confused with deletions as a cause of 21-hydroxylase deficiency.     

A long interspersed (LINE) DNA exhibiting polymorphic patterns in human genomes.

Several human DNAs digested with Kpn I restriction endonuclease released a 0.6-kilobase (kb) segment that varied in its intensity among human samples. A recombinant DNA clone (N6.4) of these 0.6-kb Kpn I segments was isolated and used to probe the genomic content and restriction cleavage pattern of homologous sequences. The hybridization patterns revealed a previously undescribed, moderately repetitive long interspersed (LINE) sequence family, which we have termed L2Hs (second LINE family in Homo sapiens). This LINE family exhibits both quantitative and qualitative polymorphisms in the human population. The content of L2Hs sequences in human genomes varies over a 5-fold range. Relative to the value for a human placental DNA, sequences homologous to the L2Hs family occur in lower amounts in gorilla DNA (approximately 20%) and even less in DNA from chimpanzees and other primates (less than 1%). Thus, the L2Hs sequences appear to have emerged only recently as a moderately repetitive sequence family in primate evolution. The observed restriction fragment length polymorphism of the L2Hs family members may reflect patterns of sequence rearrangements, amplifications, and/or deletions in human genomes.     

Exon shuffling generates an immunoglobulin heavy chain gene.

From endonuclease EcoRI partial libraries of DNAs from mouse embryo and MOPC 141, a gamma 2b-producing myeloma, clones were isolated by using a DNA fragment carrying the gamma 2b constant (C) region gene as a hybridization probe. One clone from MOPC 141 contained a heavy chain variable (V) gene and the C gamma 2b gene, as demonstrated by R-loop mapping. The V gene and C gene in this clone were separated by a 3.9-kilobase intron. The characterization of this clone as well as the embryonic clones suggest that at least two recombination events occurred to create the gamma 2b gene in MOPC 141. One of the events is analogous to the V-J joining previously demonstrated in the light chain genes, which brings the major part of the V gene next to a short coding sequence (J). The other event we refer to as "C mu-C gamma 2b switch recombination" because a portion of the intron between the V gene and C gene of the rearranged gamma 2b gene is derived from the 5' flanking sequence of the embryonic C mu gene. A model suggesting how the phenomenon of switch seen in lymphocytes may occur is presented.     

Suppression of IgE secretion from hybridoma cells in allotype congenic mice: suppression of allotype 7a by T cells in allotype b mice.

Anti-2,4-dinitrophenyl (DNP) IgE producing hybridoma B 53 when injected subcutaneously is established equally well in syngeneic BALB/c (heavy-chain allotype a) and congenic CB20 (heavy-chain allotype b) mice. However, secretion of anti-DNP IgE monoclonal antibody is greatly suppressed in CB20 mice. B 53 cells taken from the subcutaneous tumors of CB20 mice produce anti-DNP IgE in vivo in BALB/c mice and in vitro. No difference was observed in IgE production between these cells and the controls taken from BALB/c mice. The suppression of IgE production was due to T cells and/or their product(s) of CB20 mice.     

Human chromosome 7 carries the beta 2 interferon gene.

A cDNA clone (pAE20-4) corresponding to the 1.3-kilobase human beta 2 interferon mRNA was used as a probe in blot-hybridization experiments of DNA from a panel of human-rodent somatic cell hybrids containing overlapping subsets of human chromosomes. The DNA hybridization experiments showed that the human beta 2 interferon gene is located on human chromosome 7. This assignment is consistent with previous experimental data in which the expression of the translationally active 1.3-kilobase beta 2 interferon mRNA was assayed in various somatic cell hybrids. Blot-hybridization experiments using DNA from different human cell strains and cell lines reveal distinct EcoRI restriction fragment length polymorphisms of the human beta 2 interferon gene.     

DNA sequence associated with chromosome translocations in mouse plasmacytomas.

A DNA sequence that generates aberrantly rearranged immunoglobulin heavy chain constant region genes in murine plasmacytomas is shown to participate in a chromosome translocation. We have previously termed this DNA sequence NIARD for non-immunoglobulin-associated rearranging DNA. NIARD rearrangements were found frequently in murine plasmacytomas but were not detected in normal lymphocytes. These rearrangements occasionally involve the switch region of the C alpha gene. In this study, DNA samples obtained from mouse-Chinese hamster somatic cell hybrid lines were digested with various restriction endonucleases and analyzed by the Southern transfer technique with a NIARD hybridization probe. These experiments show that NIARD resides on chromosome 15 in the mouse germ line. Since NIARD is found adjacent to the C alpha gene (located on chromosome 12) in some plasmacytomas, it is apparent that a translocation involving these two chromosomes has occurred. We have proposed a rcpT(12;15) model to explain our data. The implications of NIARD rearrangements for malignant transformation are discussed.     

Non-immunoglobulin-associated DNA rearrangements in mouse plasmacytomas.

We have characterized a class of DNA rearrangements in plasmacytomas. These recombination events involve a DNA sequence whose origin is outside of the locus of the heavy chain constant region genes, CH. Therefore, we choose to refer to this sequence as non-immunoglobulin-associated rearranging DNA (NIARD). We have isolated two abortively rearranged C alpha genes, generated by NIARD events from the alpha-producing J558 myeloma. Restriction endonuclease maps of these sequences reveal two possible recombination sites in NIARD that are separated by approximately 6.5 kilobase pairs of DNA (defined as 5' and 3' sites). A NIARD rearrangement occurs in 15 out of 20 plasmacytomas tested, including gamma 3-, gamma 1-, gamma 2b-, gamma 2a-, and alpha-producers, but this event usually does not involve a CH switch (S) region. In fact, only S alpha appears to accept NIARD. However, NIARD did not undergo a rearrangement in eight IgA-producing hybridomas tested. One germ-line copy of NIARD (a 22-kilobase pair EcoRI fragment) is retained in all plasmacytomas. NIARD does not appear to possess repetitive DNA sequences homologous to S mu or S alpha. We discuss the possible role and implications of NIARD-like sequence rearrangements in allelic exclusion and chromosomal translocation events in plasmacytomas.     

Characterization of a Q subregion gene in the murine major histocompatibility complex.

We have used restriction enzyme digests, Southern blot analysis, and gene transfer experiments to identify a class I gene in the Q subregion of the murine major histocompatibility complex. By comparisons of class I genes from Q congeneic strains, five restriction fragment length polymorphisms were identified. Further studies of mutant (Qa-2-) and wild-type (Qa-2+) BALB/c sublines indicated that at least part of the structural or regulatory gene controlling a Q subregion antigen resides on a 3.7-kilobase Xba I DNA fragment and is absent in all tested Qa-2- strains. The spontaneously occurring Qa-2- BALB/cBy mutant appears to have an extensive deletion in this region. The identity of this gene was confirmed by gene transfer experiments as well as by the use of a single-copy probe.     

The ovalbumin gene: Cloning of the natural gene

The structural ovalbumin DNA sequences are not contiguous and are separated by multiple "intervening regions" in native chicken DNA. EcoRI, a restriction endonuclease that does not cleave the structural ovalbumin DNA sequences, digests the natural ovalbumin gene into three distinct fragments of 2.4, 1.8, and 9.5 kilobase pairs in length by cleaving within these "intervening regions." The 2.4-kilobase pair fragment contains only about 450 nucleotide pairs of coding sequence, with the rest being intervening sequences. This DNA fragment was cloned in bacteria by using the certified EK2 vector lambdagtWES.lambdaB after enrichment from total EcoRI-digested chicken DNA by a combination of RPC-5 column chromatography and preparative agarose gel electrophoresis. Five out of approximately 20,000 recombinant phage plaques were capable of hybridizing with a (32)P-labeled Hha I fragment of a recombinant plasmid pOV230 containing the entire structural ovalbumin gene. DNA amplified in these recombinant phages, lambdagtWES.OV2.4, was shown to contain the same restriction endonuclease cleavage sites as in the 2.4-kilobase pair EcoRI fragment previously determined by restriction mapping of total genomic chicken DNA. The intervening sequences were allowed to hybridize with excess total chicken DNA and oviduct nuclear RNA after nick-translation. They were found to be unique chicken DNA sequences, and appeared to be transcribed in their entireties during gene expression. Like the structural gene sequences, the expression of the intervening sequences is also inducible by steroid hormones.     

Partial deletion of beta-globin gene DNA in certain patients with beta 0-thalassemia.

We have used restriction endonuclease mapping of cell DNA to investigate the structure of the beta-globin gene in beta-thalassemias. Among 17 individuals with beta +- and beta 0-thalassemia, we observed three patients of Indian origin with beta 0-thalassemia whose DNA revealed a consistent mapping abnormality. In one beta allele in each diploid cell, 0.6 kilobase of DNA was deleted from beta-specific Pst I and Bgl II restriction fragments. This deletion involved 3' beta-globin gene sequences and eliminated the EcoRI site normally present at codons 121/122, but it did not extend to the BamHI site at codons 98--100 on the 5' side of the 0.90-kilobase intervening sequence normally present in beta-globin genes. Partial beta-globin gene deletion appears, therefore, to be a primary molecular defect seen in certain patients with beta 0-thalassemia.     

Cloning of human immunoglobulin epsilon chain genes: evidence for multiple C epsilon genes.

An active human epsilon chain gene was cloned from a phage library containing partial EcoRI digests of IgE-producing myeloma DNA, using the human JH (joining) gene fragment as a probe. The epsilon chain gene clone was identified by partial nucleotide sequence determination. The germ-line constant region gene of the epsilon chain (C epsilon gene) was cloned from a human fetal liver DNA library, using the cloned epsilon chain gene as a probe. Comparative studies on the human and mouse germ-line epsilon chain genes revealed that the switch (S) sequence is more conserved than the coding sequence. Restriction endonuclease BamHI digestion of human DNA produced three C epsilon fragments of 3.0, 6.5, and 9.2 kilobases, which were named C epsilon 1, C epsilon 2, and C epsilon 3 genes, respectively. We found the three C epsilon gene fragments in all of the human DNA preparations from eleven individuals. The C epsilon gene expressed in the myeloma was identified as the C epsilon 1 gene. Because the C epsilon 2 gene is deleted from the myeloma DNA, the order of the C epsilon genes is likely to be 5'-C epsilon 2-C epsilon 1-C epsilon 3-3', assuming that all the C epsilon genes are on chromosome 14. The germ-line C epsilon 3 gene was also cloned from the myeloma DNA. Characterization of the C epsilon 3 gene revealed that it does not have the S region, suggesting that it might be a pseudogene.     

Molecular cloning of translocations involving chromosome 15 and the immunoglobulin C alpha gene from chromosome 12 in two murine plasmacytomas.

Expression of IgA by plasmacytomas occurs as a result of a DNA rearrangement that brings the variable region gene, VH, a few kilobases 5' to the constant region gene, C alpha. In this study, we show that the allelic nonexpressed C alpha gene also is rearranged in most plasmacytomas. Cloning, restriction mapping, heteroduplex analyses, and sequence analyses of the nonproductively rearrange C alpha genes from two plasmacytomas, M603 and M167, have demonstrated that the nonproductive rearrangement occurs within the alpha switching region, S alpha. In each case, the same DNA sequence has been joined to the 5' side of C alpha and we have termed this DNA "NIRD" (for nonimmunoglobulin rearranged DNA). Southern blotting analyses of genomic DNAs from various IgG-, IgM-, or IgA-producing plasmacytomas suggest that NIRD is rearranged in almost all plasmacytomas. However, NIRD rearranges to the S alpha region only in IgA-producing cells, not in IgM or IgG producers. Cytogenetic evidence has shown that T(12;15) translocations are common in murine plasmacytomas. Immunoglobulin heavy chain genes are located on chromosome 12, and the translocation breakpoint in plasmacytomas occurs near the immunoglobulin genes. NIRD has been mapped to chromosome 15 by Southern blotting analysis of mouse-hamster cell lines, suggesting that the nonproductively rearranged C alpha clones represent the T(12;15) translocations identified cytogenetically. Therefore, we have identified a region of DNA on chromosome 15 that is commonly rearranged in transformed mouse lymphocytes. We speculate on the significance of NIRD in neoplastic transformation of mouse lymphocytes.     

Cloned restriction/modification system from Pseudomonas aeruginosa.

DNA fragments from Pseudomonas aeruginosa carrying the PaeR7 restriction/modification genes have been cloned in the plasmid vector pBR322 and propagated in Escherichia coli. A subclone (pPAORM3.8) has been constructed that contains the complete restriction/modification system on a 3.8-kilobase DNA fragment. Digestion of the pPAORM3.8 plasmid with nuclease BAL-31 has yielded two types of clones. One type contains an active methylase gene but no active endonuclease gene; such clones will modify the DNA but not restrict the growth of incoming phage in vivo. The second type contains an active endonuclease gene but no active methylase gene, as judged both by in vivo tests and by the activity of the cell extracts in vitro. Although extracts of cells containing these plasmids display restriction endonuclease activity, these bacteria are unable to restrict the growth of incoming phage. Furthermore, chromosomal and phage DNA isolated from these host cells are not protected against cleavage by PaeR7 in vitro. The properties of PaeR7 endonuclease and methylase enzymes have also been examined. The PaeR7 restriction endonuclease recognizes and cleaves the sequence C decreased T-C-G-A-G, as does Xho I. However, there exists a canonical Xho I site at 26.5% on the adenovirus 2 genome which is totally refractory to PaeR7 cleavage but is cut by Xho I. Under conditions of low salt, high glycerol, and high enzyme concentrations, a "PaeR7" activity is found that is similar to that observed for EcoRI. Finally, evidence is presented that the PaeR7 methylase modifies the adenine residue within the recognition sequence.     

Prevention of autoimmune insulitis in nonobese diabetic mice by expression of major histocompatibility complex class I Ld molecules.

Nonobese diabetic (NOD) mice spontaneously develop a T-cell-mediated autoimmune disease that is similar in many respects to insulin-dependent diabetes mellitus in humans. NOD mice were shown to express major histocompatibility complex class I Kd and Db antigens. To examine the possible involvement of major histocompatibility complex class I molecules in the development of autoimmune insulitis, we attempted to express a different type of class I molecule in NOD mice by crossing C57BL/6 mice transgenic for the class I Ld gene with NOD mice. The backcross progeny expressed the Ld antigen on the peripheral blood lymphocytes at a level comparable with that of the BALB/c mice. The cell surface expression of endogenous class I and class II antigens on the peripheral blood lymphocytes was not affected. Analysis of these mice revealed that the expression of the class I Ld antigen significantly reduced the incidence of insulitis at 20 weeks of age. In situ hybridization of a biotinylated probe on mouse chromosomes showed that the Ld transgene was located in the E area of chromosome 6 with which no genetic linkage to insulin-dependent diabetes mellitus was demonstrated. These results suggest that the NOD-type class I molecules are involved in the development of insulitis in NOD mice.